Hardware and Network Troubleshooting · Topic 3 of 7
Troubleshooting Drives and RAID
Diagnose failing hard drives and SSDs, boot errors, slow reads and writes, clicking noises and degraded RAID arrays, with a safe order of recovery steps.
- 6 min
- Clicking, grinding or buzzing from an HDD means mechanical failure; stop using it.
- SMART warnings mean the drive may soon fail.
- "Boot device not found" can come from the boot order, boot files or a failed drive.
What it is
Drives hold the operating system and all user data, so a drive fault can stop a computer and put data at risk. Redundant array of independent disks (RAID) groups drives to improve speed or survive a failure, and it has its own fault patterns.
The first rule is to protect data. Back up or copy what you can before you run repairs that might make things worse.
How it works
Symptoms of drive problems
| Symptom | Likely causes |
|---|---|
| Read or write failures, file errors | Failing drive, bad sectors, file system damage, loose cable |
| Slow performance | Failing drive, full drive, fragmentation on a hard disk drive (HDD), heavy background activity, not enough memory |
| Loud clicking, grinding or buzzing | Mechanical failure inside an HDD; stop using it at once |
| Drive not recognised | Loose data or power cable, wrong firmware setting, dead drive, failed controller port |
| Missing drive in the operating system | Drive not initialised, no drive letter, hidden partition, file system problem |
| Boot device not found, or operating system not found | Wrong boot order, damaged boot files, failed drive, loose cable |
| Extended read or write times | Bad sectors, failing drive |
| Self monitoring warnings | The drive reports problems through self monitoring, analysis and reporting technology (SMART) |
| Data lost or corrupted | Failing drive, sudden power loss, malware |
SMART is a status feature built into drives that records health data such as errors and temperature. Tools and the firmware can read it and warn that a failure is likely.
Low IOPS
Input/output operations per second (IOPS) counts how many reads and writes a drive or array completes each second. Low IOPS shows up as slow databases, slow virtual machines and long waits when many users read or write at once. Check these causes in turn:
- A failing drive that retries reads, shown by SMART warnings or errors.
- A degraded or rebuilding array, which is slower until it is healthy again.
- A RAID level that suits the workload badly, since parity levels such as RAID 5 write more slowly than RAID 10.
- An SSD that is nearly full, or a hard drive doing the work of an SSD.
- A damaged cable, or a port that has dropped to a lower speed.
Hard drives and SSDs
A mechanical hard drive that clicks or grinds is failing physically. Power it off and send it to data recovery if the data matters, because more use can destroy the platters.
A solid state drive (SSD) has no moving parts, so it fails differently. It may disappear suddenly, show errors or become read only.
SSDs also wear after heavy writing. Check firmware updates and the health tool of the maker.
Steps for a drive that is not seen
- Check power and data cables at both ends, and try another cable or port.
- Look in the firmware setup to see if the drive is detected.
- Check the boot order if the issue is booting.
- Check in the operating system's disk tool for an uninitialised disk or a missing letter.
- Try the drive in another computer.
- Replace the drive if it is failing, then restore data from backup.
A drive that shows in the firmware but not in the operating system is often a software problem, such as a missing drive letter or a damaged file system. A drive missing from both points to hardware, cables or the drive itself.
Boot errors
"Operating system not found" or "no boot device" can mean a wrong boot order, a removed boot drive, damaged boot files or a dead drive. Remove any unneeded removable drive or disc, check the boot order, and then test the drive.
RAID problems
| Symptom | Likely causes |
|---|---|
| RAID not found | Controller disabled, cables loose, array settings lost |
| Array degraded | One drive failed, but data is still available |
| Array failed | Too many drives failed for the level to survive |
| Rebuild is slow | Large drives, heavy use during rebuild |
| Drive keeps dropping out | Bad cable, bad port, failing drive |
When a protected array is degraded, replace the failed drive promptly with a matching or larger one, and let the array rebuild. Another failure during the rebuild can mean data loss, so ensure that backups are current.
For a RAID 0 array, one failed drive means the array is lost, so you must restore from backup. Check the controller firmware and the system's event logs for hints, and do not reinitialise an array that holds data, since that can erase it.
Drive and RAID indicator lights
Drive bays and RAID controllers have small lights that show state. The colours differ between makers, so confirm in the manual, but the general pattern is below.
| Light | Usual meaning |
|---|---|
| Off | No power, no drive or the drive is not detected |
| Steady or flickering green | Healthy, with flicker for activity |
| Amber or red | Fault or predicted failure |
| Blinking pattern | Rebuild, identify or locate function |
Check the lights and the controller utility before you pull a drive. Removing the wrong drive from an array that is already degraded can fail the whole array.
Audible RAID alarms
Many RAID controllers and servers beep when a drive fails or the array is degraded. The alarm can be silenced from the controller utility or management software, or with a mute button on some units. Silencing it does not fix anything, so still identify the failed drive, replace it and let the array rebuild.
After the repair
Restore data from backup, test the system, and set up monitoring for SMART alerts and RAID status so you learn about the next failure sooner.
What to remember for the exam
- Clicking, grinding or buzzing from an HDD means mechanical failure; stop using it.
- SMART warnings mean the drive may soon fail.
- "Boot device not found" can come from the boot order, boot files or a failed drive.
- A drive seen in firmware but not in the operating system is often a software or file system issue.
- A degraded RAID array still works; replace the failed drive and rebuild.
- RAID 0 has no redundancy, so one failed drive loses the array.
- Back up before repairs.
- RAID is not a backup.
Common mistakes
Running repair tools on a physically failing drive. Extra use can finish it off. Copy data first or use recovery services.
Reinitialising a RAID array. It can erase the data.
Ignoring a degraded array. A second failure can lose everything.
Blaming the drive when the cable is loose. Reseat and swap cables first.
Skipping backups. A repair is only safe when the data is safe.
Check yourself
What does a clicking noise from a hard drive usually mean?
The drive has a mechanical failure and should be powered off and its data recovered quickly.
What state is a RAID 5 array in after one drive fails?
Degraded; it still works, but it needs a replacement drive and a rebuild.
What does SMART report?
Health data from the drive, such as errors and temperature, which can warn that a failure is likely.
Topic 3 of 7 in Hardware and Network Troubleshooting
Test yourself on Hardware and Network Troubleshooting
The full CompTIA A+ 220-1201 question bank has 856 questions, with timed practice exams and an explanation for every answer.
This study guide is independent and is not affiliated with, endorsed by, or sponsored by CompTIA. Vendor names, certification names and exam codes are trademarks of their respective owners and are used for identification only.